Improved position control of shape memory alloy actuator using the self-sensing model

Abstract

A shape memory alloy (SMA) actuator is driven by applying electrical current (or heat) to the actuator, and its electrical resistance is altered by its deformation (or strain). The position of the SMA actuator can be controlled by referring to its electrical resistance, without extra sensors, if the resistance is used as the “feedback signal” in the control system. It is called “self-sensing capability.” However, the relationship between electrical resistance and strain of the SMA actuator inevitably exhibits a hysteresis gap, increasing the control errors. Therefore, we proposed a new self-sensing-based position control model for the SMA actuator to reduce the negative effect of the hysteresis gap and improve the accuracy of position control. The relationship between electrical resistance and strain of the SMA actuator during the application and removal of heat was first constructed. Then, we proposed a control algorithm adopting the self-sensing model considering the relationships during the heating and cooling stages. We estimated that the total mean absolute errors (MAEs) of the position control for multi-step and sinusoidal wave inputs were 3.798% and 6.380%, respectively. They were correspondingly improved by 0.649%p and 2.478%p compared to the MAEs estimated from the conventional self-sensing control algorithm. The proposed self-sensing-based position control algorithm contributes to the improvement in accuracy of the position control when the SMA actuator can be self-controlled.